Types of randomized clinical trials published in PubMed over 40 years

Abstract

Introduction: there are several variations of randomized clinical trials. Trials can be classified by design as parallel, cross-over, factorial, assignment by group, n-of-1, paired, withdrawal, adaptive and pragmatic; and by purpose as superiority, non-inferiority and equivalence. Given this heterogeneity, the Consolidated Standards of Reporting Trials (CONSORT) were introduced in 1996 to provide a guideline for reporting randomized clinical trials.

Objective: to describe the publication tendency of the various types of randomized clinical trials over 40 years, with reference to the publication of CONSORT and its extensions.

Methods: the PubMed tool was used to search for randomized clinical trials published between 1979 and 2018, classifying them according to the varieties described.

Results: a total of 472,114 published articles were found; 90.2% did not report the type of design and 98.2% did not report the purpose. Among the articles that reported the variety of randomized clinical trial, the predominant design was cross-over (5.9%), followed by parallel groups (2.34%); while the most common purpose was superiority (0.84%). After the launch of CONSORT, there was an increased proportion of articles published with the following designs: parallel groups; difference in proportions 1.89 95% CI (1.1-2.7); paired 1.07 95% CI (0.2-1.9); and pragmatic 4.73 95% CI (4.4-5.1); and for the purpose of non-inferiority 5.97 95% CI (5.6-6.3).

Discussion: most articles on randomized clinical trials do not mention their type in the title and abstract. The proportion of articles that did, increased slightly after CONSORT was published.